---
schema_version: '1.0'
id: news-20260710-6ab889
url: https://toot-books.pages.dev/blog/glm-5-2-vat-benchmark
url_hash: 6ab889a6e5472af422736a8c6d86658d620d94f4ba0e781ab95e4139e93f8480
canonical_url: https://toot-books.pages.dev/blog/glm-5-2-vat-benchmark
source: HackerNews
category: news/tech
category_raw: tech
region: usa
tags:
- it_general
- it_tech
- it_ai
lang: en
published_at: '2026-07-09T18:29:41Z'
fetched_at: '2026-07-10T00:35:26Z'
updated_at: '2026-07-10T00:37:00Z'
status: published
content_hash: null
license_note: link-only
summary: null
summary_source: null
summary_en: null
entities:
- name: IT_GENERAL
  type: concept
- name: IT_TECH
  type: concept
- name: IT_AI
  type: concept
- name: GLM 5.2
  type: artifact
- name: human
  type: person
- name: Bookshop.org
  type: organization
- name: human book keeper
  type: UNKNOWN
key_facts: []
related: []
related_auto:
- name: TLS_CERTIFICATES
  type: content
  weight: 1.0
- name: Opus
  type: artifact
  weight: 1.0
- name: GPT
  type: artifact
  weight: 1.0
- name: AI
  type: concept
  weight: 1.0
- name: Kobo eReader
  type: artifact
  weight: 1.0
title: GLM 5.2 is nearly as accurate as a human book keeper
---

# GLM 5.2 is nearly as accurate as a human book keeper

## TL;DR
（要約なし・原文はリンク先を参照 / No summary available; see the source link.）

## Key Points
- it_general / it_tech / it_ai

## Details
(本文なし。リンク先参照)

## Source
元記事: [GLM 5.2 is nearly as accurate as a human book keeper](https://toot-books.pages.dev/blog/glm-5-2-vat-benchmark) — published 2026-07-09T18:29:41Z